The basic idea behind FLEXI-cross is to augment and enhance border-checking practices with a plethora of innovative — and even portable — solutions, based on technologies such as augmented reality, “tethered” UAVs, the blockchain, and “multimodal” biometric technologies (including face recognition).
Writes the project’s Cordis page: “The FLEXI-cross project aims to increased security and reliability of EU border checks for people and goods, while enhancing and improving the border management capabilities through the development, deployment and validation of a toolkit of innovative border-checking solutions, in real operational environments, addressing road, rail and port borders. The resulting flexibility and dynamicity of border check planning will offer novel capabilities such as dynamic deployment of check-points and support via mobile applications for border personnel, while guaranteeing high level of security, privacy of personal data and protection of people’s fundamental rights.”
It is worth stressing, from the last sentence, that a “high level” — rather than full — protection of fundamental rights is promised.
A project overview is also contained in a set of slides.
According to the official description on Cordis, FLEXI-cross solutions will feature the following functionalities:
“i) anti-trafficking and anti-smuggling protection via predictive risk assessment of vehicle and people (“enabling specific patterns detection”, according to the project’s site, ndr),
ii) enhanced border security through portable biometric based checks (which include fingerprint and facial recognition, according to the project’s site, ndr),
iii) secure person verification through real-time multi-source cross-referencing,
iv) flexible, fast and cost-effective deployment of ad-hoc Border Check Points (BCP),
v) secure, private and traceable sensitive / personal data exchange based on blockchain technology and
vi) increased safety and improved experience for border-personnel based on advanced Human Machine Interfaces and enhanced situational awareness via Augmented Reality (AR).”
The logics behind all of this is to save time and money, through “more efficient border checking procedures”, thanks to “optimisation (flow management, active planning), based on real-time data, historical data, available resources and traffic estimations”, writes the project site.
Newsletter #2 argues that “AR and Situational awareness” are “a necessity for the border security”: “AR Situational Awareness integrates augmented reality technology into border security systems, providing real-time, contextual information to enhance security personnel’s understanding of the environment.”
In FLEXI-cross, more specifically, “The AR Situational Awareness module in the FLEXI-cross toolkit provides holographic visualisation of pertinent and critical information, such as suspicious people and dangerous goods/materials. HOLO (i.e., Hololight, a German firm that “specializes in immersive software and technology for the enterprise market”, ed.) will supply the AR visualisation software and hardware for the 3D AR Situational Awareness. This module will be demonstrated in a use case at Galati Port, specifically during port border checks. The advanced AR Situational Awareness module is set to be utilised in five different scenarios within this use case:
1. Identification of a mismatch between the documentation provided and the individual,
2. Fraudulent abandonment of a ship by a crew member during border control,
3. Fraudulent boarding of a docked ship by a foreign person,
4. Detection of suspected dangerous materials on the boarded ship,
5. Detection of an unauthorised individual on the vessel”.
Biometrics, surveillance, AI, and augmented reality are deeply intertwined in FLEXI-cross. In fact, “HOLO’s AR Situational Awareness module will be combined with partner ICCS’s face verification and identification modules, as well as intelligent surveillance modules. Additionally, it will integrate acoustic and radiation device modules from BEIA partners. This integrated system is set to be deployed in the Galati port for use by end-user partners to optimise border security management.”
Some more details on AR Situational Awareness can be found on a blog post.
Newsletter #3 details how the blockchain will be also implemented in FLEXI-cross “to improve border security, specifically targeting cargo identification and management at the Romania-Moldova borders in Giurgiulesti” — described as a “crucial development, set for mid-summer 2024 implementation”, one that “marks a significant advancement in modern border management, fostering trust and collaboration across borders”.
For example, “One major advantage that blockchain offers in the use case at the Romanian – Moldavian rail borders, is the potential for cross-border collaboration for both nations. Blockchain’s decentralized nature can foster better collaboration between Romanian and Moldovan border control authorities by providing a shared and trusted platform for data exchange. This can lead to more aligned procedures and faster processing times.”
A blogpost on FLEXI-cross’s website reminds that Dr. Giuseppe Vella, as FLEXI-cross Project Coordinator, participated in the fourth Project to Policy Seminar (“an annual event jointly organized by DG HOME and REA”), in which he “engaged with 2 similar projects, iSEAMORE and ODYSSEUS (…). Together, they discussed the way forward for creating synergy between the 3 projects to increase the public awareness, collaboration and to disseminate the project results and communicate the organisation meetings or events organised by the 3 projects, aiming to attract more targeted audience. In addition, and during the second day of the seminar, FLEXI-cross has participated in an open session with other projects (…) and (…) came in contact with the project coordinators of TENSOR and TENACITY that manages biometric authentication and algorithms respectively for forensic purposes and for fighting crimes committed during cross borders.”
Project objectives include “upgrading existing infrastructure and legacy systems from consortium partners and previous EU-funded projects” (these are however not specified).
In March 2024, a “synergy workshop” with projects ODYSSEUS, TENSOR and EITHOS (a project developing “a novel Identity Theft Observatory System”) was also held.
In the redacted Grant Agreement we managed to obtain through access to information requests to the Research Executive Agency (REA), we also found that “most” Consortium partners can boast “proven success records in National and H2020 Projects”, with some even considered “highly experienced with significant resources and extensive participation in FP6, FP7 and H2020 projects with valid roles”.
Crucially, the solutions developed within FLEXI-cross all build on specific technological achievements by previous EU-funded projects, to ensure a “high Technological Readiness Level”, and therefore potentially lead to a quick actual deployment.
The GA details some of them:
“- new modalities for biometric based border-checks (BES-06-2015 PROTECT project),
– acceptance of no gate crossing points solutions (SEC-18- BES-2017 PERSONA project),
– risk-based screening at border crossing (SEC-15-BES-2017 TRESSPASS project),
– technologies to enhance border and external security (SU-BES02-2020 BORDERSENS project),
– “zero-touch” border crossing technology for Connected and Automated mobility corridors (ICT-18-2018 5G-MOBIX project), and more.”
Specific links are also drawn with EFFECTOR (for its “multi-layered data fusion platform”), BORDERUAS (to “utilize the knowledge related to UAVs use in border surveillance that was acquired from this project”), ANDROMEDA (for the knowledge developed on “tethered” UAVs) and, noticeably, the controversial iBorderCtrl project.
Writes the GA: “Links to iBorderCTRL: FLEXI-cross will employ experience and knowledge such as the mobility concepts, processes designed to reduce cost/time spent per traveller at the crossing station, as well as technologies used will be considered.”
Deliverable D1.2 writes provides “a detailed description of the three use cases of the FLEXI-cross project, each of which serves as a platform to validate the solutions produced during the project”.
More specifically, these concern: “1) port-based crossing at the Romanian Port of Galati, 2) vehicular road-based crossing at the Greek-Bulgarian border and 3) rail-based crossing at the Romanian-Moldovan border”.
Both the developed toolkit and the use cases were collaboratively designed according to the “needs of relevant border authorities”.
Operational and technical details are also included, for example letting us know that
1) The Romanian use case includes IoT components, “novel high-power acoustic through-the-wall” sensors, “AI/ML tools for risk assessment based on acoustic sensor input”, “3D AR situational awareness” (meaning that “AR smart glasses will be used to assist border control personnel with the identification of people that have been detected and identified as suspicious or warranting attention”), a “Surveillance-Crowd Behaviour Analysis” platform and a “Face Verification Identification mobile application”;
2) the Greek-Bulgarian use case scenarios include that of a “Truck carrying hidden migrants”, to be tackled with “Machine Vision analytics via infrastructure cameras or fixed cameras in tethered drones”;
3) in the Romanian-Moldovan use case, the creation of “a software platform that allows relevant authorities to link parents’ and their children’s biometric identification data (including multimodal biometry using face/voice) and travel documents, also usable by border control authorities and featuring delegation of custody rights using blockchain” to fight children trafficking.
Final results for these use case scenarios are not available to the public yet, at the time of writing. However, the objective is to “showcase the maturity and business readiness of the solutions”.
Also, it is interesting to note these objectives will be obtained “by upgrading existing infrastructure and legacy systems from consortium partners and previous EU-funded projects to support the thorough testing of the use cases in road, port and rail environments”.
Only 4 deliverables are publicly available on Cordis at the time of writing — whereas none is available on the project’s website, where no publications are featured so far either.
In general, from what we can understand of it at this point, the project seems to be strongly informed by a solutionist logic (i.e., one that is focussed on the necessity of providing an optimal solution for maximum efficiency through technology).
This is clearly shown for example in how D1.2 describes the use case for the Romanian pilot: “it is urgently necessary to increase the level of security at Romania’s borders by implementing technical solutions that allow faster and more accurate detection of high-risk persons and types of goods considered dangerous (e.g., waste, weapons, drugs, etc.) by using the latest generation of technological solutions.” (emphasis added, ed.)
A “crimmigration” framework (i.e., one that fundamentally conflates crime and migration, ed.) is also evident, including in the framing of the policy workshops attended by FLEXI-cross representatives: “Participants were organized into thematic areas such as Fighting crime and terrorism & Infrastructure (FCT/INFRA), Strengthened Security Research and Innovation (SSRI), Disaster-resilient societies (DRS), and Border management (BM).”
Furthermore, portable biometrics devices in the context of migration — were they to be actually deployed at EU borders — should be included among “high-risk” systems in the AI Act according to the #ProtectNotSurveil coalition (cfr. 5.A here). Predictive risk assessment is also extremely problematic in the context of border management and surveillance — and should in certain cases be banned according to the same coalition.
This project, like many others, also aims at updating and expanding border surveillance facilities and tools: but how much can these be further updated, augmented and expanded before they become structurally incompatible with “EU values” (fundamental rights, the rule of law and democratic principles themselves)? Is such a tipping point even conceivable? And is there more generally an end to the ever-expanding reach of border (and “prefrontal”) surveillance? Apparently, not for FLEXI-cross, which among its aims has the following: “Accelerate the adoption of the proposed dynamic, flexible and mobile technological solutions by the border authorities, and to maximise the resulting impact on daily border management and operations, in close liaison with existing projects and initiatives”.
Interestingly, the project promises “a high level” of fundamental rights protection, and yet no input/co-creation procedure is envisioned (still) to include the voices of those affected by the tech being developed.
The project is purely aimed at enforcement agencies: “The design and architecture of the implemented solutions will be based on the analysis of the technical and service requirements of the EU border authorities (e.g., Frontex) and will offer seamless integration with existing systems and devices currently utilised by them” (Newsletter #1).
Lastly, the logic behind the project is one of totalizing surveillance, to ultimately obtain “seamless” and secure border crossings. Writes for example D1.2 that the Greek-Bulgarian use case would encompass “An infrastructure camera or camera fixed on tethered drone deployed in order to observe through the whole length of the queue and provide real time streaming for machine vision analytics on plates’ number recognition, vehicles’ queuing number, passengers’ suspicious behaviour (e.g., U-turn deviating from the queue). On board sensors and remote connectivity to vehicles’ OBU (e.g., for trucks) can provide additional information on the vehicles’ travel history, stops, engine and cargo status, potential contraband and/or smuggling attempts and more. A Pre-registration mobile application used by passengers and vehicles can be used, to minimize waiting times at the actual borders”.
As an important final methodological note, all the FLEXI-cross deliverables we obtained — namely, its Grant Agreement and deliverables D1.2 and D4.5 — were still “subject to review and approval of REA”, wrote the EU Commission agency in response to our access requests, and should therefore “be considered as drafts and the final version might differ, in some cases even significantly.”
